Gliarin and Macrolin, Two Novel Intermediate
Filament Proteins Specifically Expressed in
Sets and Subsets of Glial Cells in Leech
Central Nervous System
Yingzhi Xu,1 Brian Bolton,1 Birgit Zipser,2 John Jellies,3 Kristen M. Johansen,1
Jørgen Johansen1
1
Department of Zoology and Genetics, 3156 Molecular Biology Building, Iowa State University,
Ames, Iowa 50011
2
Department of Physiology, Michigan State University, East Lansing, Michigan 48824
3
Department of Biological Sciences, Western Michigan University, Kalamazoo, Michigan 49008
Received 6 January 1999; accepted 19 February 1999
ABSTRACT: Using monoclonal antibodies, we
have identified two novel intermediate filament (IF) proteins, Gliarin and Macrolin, which are specifically expressed in the central nervous system of an invertebrate.
The two proteins both contain the coiled-coil rod domain
typical of the superfamily of IF proteins flanked by
unique N- and C-terminal domains. Gliarin was found
in all glial cells including macro- and microglial cells,
whereas Macrolin was expressed in only a single pair of
giant connective glial cells. The identification of Macrolin and Gliarin together with the characterization of the
strictly neuronal IF protein Filarin in leech central ner-
Intermediate filaments (IFs) are cytoskeletal proteins
that constitute a diverse multigene family that contains more than 50 different IF genes (Albers and
Fuchs, 1992). The IFs are differentially expressed
within nearly all cell types and serve as mechanical
integrators of the cytoplasm that function to resist
mechanical stress (Fuchs and Cleveland, 1998). DeCorrespondence to: J. Johansen
Contract grant sponsor: NIH; contract grant number: NS 28857
Contract grant sponsor: NSF; contract grant number: 9724064
Contract grant sponsor: Howard Hughes Medical Institute Education Initiative
Contract grant sponsor: Hatch Act and State of Iowa
© 1999 John Wiley & Sons, Inc. CCC 0022-3034/99/020244-10
244
vous system demonstrate that multiple neuron- and glial-specific IFs are not unique to the vertebrate nervous
system but are also found in invertebrates. Interestingly,
phylogenetic analysis based on maximum parsimony indicated that the presence of neuron- and glial cell–
specific IFs in coelomate protostomes as well as in vertebrates is not of monophyletic origin, but rather
represents convergent evolution and appears to have
arisen independently. © 1999 John Wiley & Sons, Inc. J Neurobiol
40: 244 –253, 1999
Keywords: glial cells; intermediate filaments; nervous
system; leech; phylogenetic analysis
spite their diversity all IF proteins share common
structural features which include an a-helical rod
domain defined by regions of heptad repeats in which
the first and fourth residue usually are hydrophobic or
nonpolar (Steinert and Roop, 1988). The rod domain
is flanked by variable N- and C-terminal domains
which are responsible for a major part of the structural
heterogeneity of IF proteins (Steinert and Roop,
1988). Based on similarities in sequence structure and
intron placement vertebrate IFs have generally been
divided into six classes (Steinert and Roop, 1988;
Lendahl et al., 1990; Albers and Fuchs, 1992): type I
and II keratins, type III cytoplasmic IFs, type IV
neurofilaments, type V nuclear lamins, and type VI
Glial Cell–Specific Intermediate Filaments
nestins. In addition to the six types of vertebrate IFs,
an increasing number of invertebrate IFs have also
been cloned and characterized (Weber et al., 1988,
1989; Szaro et al., 1991; Tomarev et al., 1993; Dodemont et al., 1994; Johansen and Johansen, 1995;
Bovenschulte et al., 1995). Interestingly, invertebrate
IFs share a feature with the nuclear lamins of having
six heptad repeats in the rod domain which are not
found in vertebrate cytoplasmic IFs, raising questions
as to the evolutionary history of IF proteins (Osborn
and Weber, 1986; Weber et al., 1988; Steinert and
Roop, 1988).
We were particularly interested in IFs’ role and
distribution within the invertebrate nervous system
and whether IFs specific to different components of
the central nervous system such as neurons and glia
exist. Previously only two types of such IFs, squid
brain IF (Szaro et al., 1991; Way et al., 1992) and
Filarin (Johansen and Johansen, 1995), have been
demonstrated in the invertebrate central nervous
system, and they are both selectively expressed by
neurons. However, in this study we show that in
addition to neurons, IFs can also be found to be
specifically expressed in sets and subsets of glial
cell types within the leech central nervous system.
The leech central nervous system consists of a head
and a tail brain and 21 similar segmental ganglia
linked to each other by connectives (Muller et al.,
1981). Each segmental ganglion contains about 400
neurons which are relatively large, some being up
to 100 mm in diameter (Macagno, 1980). In addition to microglial cells, there are four types of giant
glial cells present within the central nervous system
in each segment (Coggeshall and Fawcett, 1964;
Lüthi et al., 1994): (a) one pair of macroglial cells
located in the connectives wrap all axons traveling
in the two lateral connectives and Faivre’s nerve,
(b) six packet glial cells envelope the neuronal cell
bodies in each ganglion, (c) two large glial cells
surround axons and neuronal processes in the neuropil, and (d) two glial cells ensheathe axons in the
peripheral nerve roots. Using two new monoclonal
antibodies (mABs) first reported here, 9A8 and
1A11, as well as two previously characterized
mABs, G39 (Lüthi et al., 1994) and Lan3-13
(Flaster and Zipser, 1987; Morrisey and McGladeMcCulloh, 1988), we show that all types of glial
cells within the leech central nervous system express a novel IF protein, which we have named
Gliarin, whereas the connective macroglial cell, in
addition, specifically expresses a different IF protein which we have named Macrolin.
245
MATERIALS AND METHODS
Experimental Preparations
For the present experiments, we used two different leech
species—namely, the hirudinid leeches Hirudo medicinalis
and Haemopis marmorata. The leeches were either captured
in the wild or purchased from commercial sources. Dissections of nervous tissue were performed in leech saline
solutions with the following composition (in mM): 110
NaCl, 4 KCl, 2 CaCl2, 10 glucose, 10 Hepes, pH 7.4.
Antibodies and Ascites Production
Monoclonal antibody production essentially followed the
procedure of Zipser and McKay (1981). In short, Balb C
mice were immunized at 21-day intervals by intraperitoneal
injection with a homogenate of the entire nerve cords from
four Hirudo leeches. Four days after a boost by intravenous
injection of sodium dodecyl sulfate (SDS)-extracted nerve
cords, spleen cells of the mice were fused with Sp2 myeloma cells and a number of monospecific hybridoma lines
were established using standard procedures (Harlow and
Lane, 1988). Two of these hybridoma lines, 1A11 and 9A8,
were specific for leech glial cells and were selected for
further analysis in the present paper. In addition, three
previously reported mABs, G39 (Lüthi et al., 1994),
Lan3-13 (Flaster and Zipser, 1987; Morrisey and McGladeMcCulloh, 1988), and Lan3-8 (McKay et al., 1984; Johansen and Johansen, 1995) were used in these studies. Ascites
fluids from mAB 1A11 and G39 were obtained by injecting
four mice for each mAB intraperitoneally with antibodyproducing hybridoma cells. All the mABs used in this study
show cross-reactivity in both Hirudo and Haemopis leeches.
Immunocytochemistry
Dissected leech nerve cords were fixed overnight at 4°C in
4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4,
the connective capsules on the ventral side were opened
with fine forceps, and the ganglia xylene extracted for better
antibody penetration (Zipser and McKay, 1981). The nerve
cords were incubated overnight at room temperature in
either hybridoma supernatant or diluted ascites fluid of the
mABs 9A8, 1A11, G39, Lan3-8, or Lan3-13 containing
0.4% Triton X-100, washed in phosphate-buffered saline
(PBS) with 0.4% Triton X-100, and incubated with horseradish peroxidase (HRP)-conjugated goat anti-mouse antibody (BioRad; 1:200 dilution). After washing in PBS, the
HRP-conjugated antibody complex was visualized by reaction in 3,39-diaminobenzidine (DAB) (0.03%) and H2O2
(0.01%) for 10 min. The final preparations were dehydrated
in alcohol, cleared in xylene, and embedded as whole
mounts in Depex mountant. The labeled preparations were
photographed on a Zeiss Axioskop using Ektachrome 64T
film. The color positives were digitized using Adobe Photoshop and a Nikon Coolscan slide scanner. In Photoshop,
the images were converted to black and white and image
246
Xu et al.
processed before being imported into Freehand (Macromedia) for composition and labeling.
Molecular Cloning and Sequence
Analysis
Ascites fluid from the mABs 1A11 and G39 and hybridoma
supernatant from the Lan3-8 and 9A8 antibodies were used
to screen a random-primed Hirudo central nervous system–
enriched cDNA lambda-ZAP II expression library (Huang
et al., 1997) essentially according to the procedures of
Sambrook et al. (1989) at a density of 30,000 plaqueforming units/150-mm plate. Positive clones were plaque
purified and in vivo excised to generate pBluescript phagemids according to the method provided by the manufacturer
(Stratagene). Several partial cDNAs were identified by each
antibody in these screens. To identify additional clones to
obtain the full sequence of the cDNAs for Gliarin, Macrolin,
and Hirudo Filarin, the same cDNA library was rescreened
using 32P-labeled fragments of the originally identified
clones. The fragments were radiolabeled using random
priming according to the manufacturer’s procedure (Primea-Gene kit; Promega) and the library screened using standard procedures (Sambrook et al., 1989).
DNA sequencing was performed using an Applied BioSystem DNA Sequencer 377A at the ISU Nucleic Acid
Sequencing Facility using commercially available universal
and reverse-sequencing primers (Stratagene) or specific
primers synthesized at the ISU DNA Sequencing and Synthesis Facility based on the determined sequences. The
nucleotide and predicted amino acid sequences were analyzed using the GCG (Genetics Computer Group Package,
Version 8, Madison, WI) suite of programs (Devereux et al.,
1984). The Gliarin, Macrolin, and Filarin sequences were
compared with known and predicted proteins in the
SwissProt and Genbank databases using the FASTA and
TFASTA programs within the GCG package. In addition, a
BLAST search was performed using the NCBI BLAST
E-mail server (Altschul et al., 1990) comparing the Macrolin, Gliarin, and Filarin sequences with SwissProt, PIR, and
GenPept databases.
Biochemical Analysis
Sodium dodecyl sulfate–polyacrylamide gel electrophoresis
(SDS-PAGE) was performed according to standard procedures (Laemmli, 1970). Electroblot transfer was performed
as in Towbin et al. (1979). For these experiments we used
the BioRad Mini Protean II system, electroblotting to 0.2
mm nitrocellulose, and using anti-mouse HRP-conjugated
secondary antibody (BioRad) (1:3000) for visualization of
primary antibody diluted in Blotto for immunoblot analysis.
The signal was developed with diaminobenzadine (0.1 mg/
mL) and H2O2 (0.03%) and enhanced with 0.008% NiCl2.
The Western blots were digitized using Photoshop software
and an Arcus II scanner (AGFA).
Phylogenetic Analysis
Alignments used to produce maximum parsimony trees
were generated with the Clustalw version 1.7 program and
initially encompassed the rod domain of the IFs analyzed.
However, in the final analysis any gaps in the resulting
alignments were removed by deleting residues corresponding to the gaps. In the analysis restricted to the cytoplasmic
coelomate invertebrate IF sequences, the initial alignment
consisted of the entire IF sequence before gap removal.
Trees were constructed by maximum parsimony using the
PAUP computer program version 3.1.1 (Swofford, 1993) on
a Power Macintosh G3. All trees were generated by heuristic searches and bootstrap values in percentage of 1000
replications (Felsenstein, 1985) are indicated on the bootstrap 50% majority rule consensus trees.
RESULTS
To obtain probes specific to different components of
the leech nervous system, panels of mABs were generated from immunizing mice with a protein homogenate obtained from whole Hirudo nerve cords. Two
of the resulting mABs, 9A8 and 1A11, showed immunocytochemical labeling restricted to glial cells in
whole-mount preparations of leech nerve cords and
ganglia (Fig. 1). While mAB 9A8 labeled all known
glial cells including microglia and the four types of
giant glial cells [Fig. 1(C)], labeling of mAB 1A11
was restricted to the connective macroglial cell [Fig.
1(A,B)]. There are only two of these cells per segment, as each cell is responsible for wrapping the
thousands of axons in each of the paired lateral connectives, and one or the other also wraps the axons of
the unpaired Faivre’s nerve. The connective macroglial cells are truly giant, as they can be up to a
centimeter long [Fig. 1(A)]. Previously, an mAB,
G39, was reported (Lüthi et al., 1994) to have a
staining pattern similar to that of mAB 9A8, whereas
the Lan3-13 antibody described by Flaster and Zipser
(1987) and by Morrisey and McGlade-McCulloh
(1988) was specific to the connective macroglial cell
as in the case of mAB 1A11. We therefore compared
the labeling of these antibodies with mABs 9A8 and
1A11 on immunoblots (Fig. 2). In this analysis, the
labeling of mAB 9A8 was indistinguishable from that
of mAB G39, while the labeling of mAB 1A11 was
indistinguishable from that of Lan3-13. mABs G39
and 9A8 both recognized a triplet of bands, with the
top band having an apparent molecular weight of 70
kD. On the other hand, the mABs 1A11 and Lan3-13
recognized a doublet, with the largest band having an
apparent molecular weight of 78 kD. The lower bands
in both cases may have represented proteolytic frag-
Glial Cell–Specific Intermediate Filaments
247
Figure 1 Monoclonal antibody labeling of glial- and neuron-specific IF proteins in whole-mount
preparations of leech nerve cords. (A) Labeling of Macrolin in the two giant connective glial cells
by mAB 1A11. The micrograph is a composite showing the labeling of the entire connective
between two ganglia (g). Scale bar 5 600 mm. (B) Higher magnification of one of the ganglia in (A)
demonstrating the labeling of Macrolin by mAB 1A11 in the connective macroglial cells at the
interphase with the neuropil. (C) Labeling of Gliarin by mAB 9A8 in a leech ganglion. All macroand microglial cells are labeled including the connective, neuropil, root, and packet glial cells. The
wrappings of the packet glial cells of some of the larger neurons are discernible. (D) Labeling of
Filarin by the Lan3-8 antibody of all neurons in a leech ganglion. In all figures the ventral side of
the nerve cord is shown with anterior to the left. Scale bar for (B–D) 150 mm.
ments. These data strongly suggest that mABs 9A8
and G39 recognize the same 70-kD antigen, whereas
mABs 1A11 and Lan3-13 both recognize a different
larger 78-kD antigen. The molecular sizes and the
immunocytochemical staining for both antigens made
them candidates for being IF proteins. Direct evidence
for this has been provided in the case of the mAB G39
antigen from which a partial amino acid sequence,
QNQQLSDYEGEISLL, was obtained by Lüthi et al.
(1994). This sequence showed 80% homology to a
stretch within the rod domain of the squid IF protein
NF60 (Szaro et al., 1991). For these reasons, and to
obtain the complete amino acid sequence of these
antigens, we screened a Hirudo central nervous system expression vector library with the glial cell–
specific antibodies.
Gliarin
The leech cDNA library was screened independently,
with both the 9A8 and G39 mABs, resulting in the
identification of several hundred partial cDNA clones
by each antibody, of which four and six, respectively,
were selected for further analysis. Subsequently, the
cDNA library was rescreened with radiolabeled nucleotide probes generated from the 59 ends of these
original cDNA clones. In this way, additional independent and overlapping cDNA clones were isolated
that encompassed the entire coding sequence. The
predicted sequence for the protein identified in both
the mAB 9A8 and G39 screen was identical and
contained 639 amino acids with a calculated molecular mass of 71,340 DD. In addition, the protein contained the exact QNQQLSDYEGEISLL sequence
previously determined from the immuno-affinity-purified mAB G39 antigen (Lüthi et al., 1994), confirming the identity of the isolated cDNA. Since the 9A8/
G39 antigen is expressed by all glia cells in the leech
nervous system, we have named it Gliarin. The complete sequence for Gliarin is shown in Figure 3, and
the inferred protein product contains all the defining
features of an IF protein (Steinert and Roop, 1988). It
248
Xu et al.
clones were identified. From rescreening with radiolabeled nucleotide probes from the 59 and 39 ends of
the original cDNA clones, additional independent and
overlapping cDNA clones were isolated that covered
the entire coding sequence. The predicted sequence
was for a protein containing 688 amino acids, which
we have named Macrolin, since it was expressed
specifically by the giant connective macroglia cells.
The calculated molecular mass of Macrolin of 79,004
DD was close to the estimate for the mAB 1A11
antigen of 78 kD based on SDS-PAGE (Fig. 2). As in
the case of Gliarin, Macrolin contains an a-helical rod
domain of 354 residues typical of IF proteins (Fig. 3)
flanked by a 142–amino acid N-terminal head domain
and a 192-residue C-terminal tail domain. The original cDNA clones identified by mAB 1A11 were also
labeled by the Lan3-13 antibody, but not by the mABs
9A8 or G39 (data not shown). These data, together
with the fact that the labeling of mAB 1A11 and
Lan3-13 on immunoblots of leech central nervous
system proteins are indistinguishable (Fig. 2),
strongly suggest that both mAB 1A11 and Lan3-13
recognize Macrolin.
Filarin
Figure 2 Immunoblots of leech central nervous system
extracts labeled with the mABs G39, 9A8, 1A11, Lan3-13,
and Lan3-8. Gliarin is identified as a 70-kD protein by the
mABs G39 and 9A8, Macrolin as a 78-kD protein by the
mABs 1A11 and Lan3-13, and Filarin as a 63-kD protein by
the Lan3-8 antibody. Molecular weight markers is indicated
to the left in kilodaltons.
has a 96 –amino acid N-terminal head domain followed by a a-helical rod domain of 355 residues (Fig.
3) which shows the typical segmentation into subdomains (coils 1A, 1B, 2A, and 2B) separated by short
nonhelical spacers (linkers L1, L12, and L2). The
coiled-coil regions conform with the repeated heptad
unit structure (a,b,c,d,e,f,g)n, where a hydrophobic or
nonpolar amino acid is usually in the a and d positions
and the other positions are occupied by either polar or
charged amino acids (Steinert and Parry, 1985). The
C-terminal tail domain consists of 188 amino acids.
Coil 1B contains the six additional heptad repeats
characteristic of nuclear lamins and invertebrate cytoplasmic IFs (Fuchs and Weber, 1994).
Macrolin
Of approximately 4.5 3 105 clones screened with the
mAB 1A11, two partial antibody-positive cDNA
We have previously identified the neuron-specific IF
protein Filarin in the leech genus Haemopis using the
Lan3-8 antibody (Johansen and Johansen, 1995). For
comparative purposes, we also screened the Hirudo
cDNA library with the Lan3-8 antibody and identified
a full-length Hirudo cDNA clone of Filarin. Hirudo
Filarin is a 597-residue IF protein (Fig. 3) with a
predicted molecular mas of 67,168 DD. Hirudo Filarin is 89% identical to the amino acid sequence of
Haemopis Filarin.
Phylogenetic Analysis
Figure 3 shows a sequence comparison of Gliarin,
Macrolin, and Filarin with each other. While the Nterminal amino acid sequences are mainly unique,
there are many scattered regions of sequence homology in the rod domains and part of the C-terminal tail
domains. The most conserved domains between Gliarin, Macrolin, and Filarin, as for all IF proteins, are at
the beginning and end of the rod domain. Macrolin
and Gliarin share 49.7% sequence identity in the rod
domain, whereas Filarin is 36.2% and 46.7% identical
in this region to Macrolin and Gliarin, respectively.
However, this level of conservation is approximately
the same as the amino acid identity to the corresponding sequences in other invertebrate IF proteins. Thus,
to determine the evolutionary relationship between
Glial Cell–Specific Intermediate Filaments
Figure 3 Alignment and comparison of the predicted amino acid sequences for Hirudo Gliarin,
Macrolin, and Filarin. Gliarin consists of 639 amino acids with a calculated molecular mass of 71.3
kD, Macrolin has 688 residues with a molecular mass of 79.0 kD, and Filarin is a 597–amino acid
249
250
Xu et al.
Macrolin, Gliarin, and Filarin and other members of
the IF protein superfamily, we constructed phylogenetic trees based on maximum parsimony. Figure
4(A) shows a consensus tree based on an alignment
with all gaps removed (leaving 274 residues) of the
rod domains of 36 IF proteins from nuclear, neuronal,
and nonneuronal cytoplasmic invertebrate IFs as well
as representative sequences from the six classes (I–
VI) of vertebrate IFs. The tree was rooted using
sequences from three pseudocoelomate invertebrate
IFs (nematode ifA3, ifC1, and cif) as an outgroup.
This outgroup was chosen based on the analysis of the
topology of individual unrooted maximum parsimony
trees. Gliarin, Macrolin, and Filarin were clearly
grouped together in a monophyletic clade constituted
by neuronal and nonneuronal cytoplasmic coelomate
protostomic IFs [Fig. 4(A)]. Interestingly, however,
all the nuclear lamins from both vertebrates and invertebrates were grouped together in a monophyletic
clade that shared a common ancestor with the monophyletic clade containing the remaining five classes of
vertebrate cytoplasmic IFs. Thus, these results suggest
that vertebrate cytoplasmic IFs may be more closely
related to the nuclear lamins than to the cytoplasmic
coelomate protostomic IFs. All the major clades in
this analysis have strong bootstrap support of .80%
[Fig. 4(A), values in italic].
Since the evolutionary relationship between the
neuronal and nonneuronal cytoplasmic coelomate
protostomic IFs was poorly resolved in the tree depicted in Figure 4(A), we realigned these protein
sequences with each other in their entirety. After all
gaps were removed from this alignment, 542 residues
remained. Figure 4(B) shows a rooted consensus tree
based on this alignment using the nematode sequence
ifA3 as an outgroup; however, the same topology was
obtained for unrooted trees. Filarin, Macrolin, and
Gliarin are clearly closely related, as they formed a
monophyletic clade together with an IF sequence
from earthworm (Bovenschulte et al., 1995). The
earthworm IF is most closely related to Macrolin;
unfortunately, its cellular expression has not been
determined. Since it formed a clade with Gliarin and
Macrolin, it would be interesting to know whether it is
also expressed by glial cells and might be a functional
homolog. Although Filarin is a neuron-specific IF, it
is clearly not a homolog of the only other neuronspecific IF to be characterized, squid brain IF (Szaro
et al., 1991). The three leech IFs were grouped in a
monophyletic clade together with the non-neuronal
cytoplasmic IFs separately from a clade formed by the
squid brain IF and a homolog recently identified in
Helix (Adjaye et al., 1995).
DISCUSSION
This study reports the cloning and characterization of
two novel IF proteins that are specifically expressed in
glial cells of the leech central nervous system. Gliarin
is a 71-kD protein present in all glial cells including
both macro- and microglial cells (Lüthi et al., 1994).
Using whole-mount and sectioned preparations, Lüthi
et al. (1994) showed that during development, Gliarin
is first discernible in 8-day-old embryos; however, at
this stage, only the anlagen for the glial cell bodies are
present. The elaborate arborizations of the glial processes progressively develop in the following days,
until in 21-day-old embryos these cells closely resemble those of adult nerve cords (Lüthi et al., 1994). In
contrast to Gliarin, Macrolin is expressed in a single
glial cell type of which there are only two per segment. These cells are large macroglial cells that envelope the thousands of axons traveling in the two
lateral nerves as well as the unpaired Faivre’s nerve
which form the ganglionic connectives. Thus, these
findings imply that the connective macroglial cells
express at least two different IF proteins. However,
whether Macrolin and Gliarin form independent
fibrils of homodimers or whether in these cells they
can form heterodimers remains to be determined. The
fact that only this type of glial cell exhibits both IFs
suggests that their dual expression may be related to
the requirement for structural integrity of these cells,
as they have to maintain extensive arborizations wrapping all axons in the entire connective. While Gliarin
and Macrolin are the first glial cell–specific IFs to be
reported on in invertebrates, the cytoplasmic IF GFAP
serves as a marker for astrocytes in vertebrates (Fuchs
and Weber, 1994).
All invertebrate cytoplasmic IFs, nonneuronal as
well as neuronal, have been found to share a charac-
protein with a predicted molecular mass of 67.2 kD. Shared amino acids between these IFs are in
white typeface outlined in black and the beginning and end of the coiled-coil rod domain is indicated
by arrowheads. Macrolin and Gliarin share 49.7% sequence identity in the rod domain, whereas
Filarin is 36.2% and 46.7% identical in this region to Macrolin and Gliarin, respectively. These
sequence data and the corresponding nucleotide sequences are available from EMBL/GenBank/
DDBJ under Accession Nos. AF101065 (Gliarin), AF101064 (Macrolin), and AF101063 (Filarin).
Glial Cell–Specific Intermediate Filaments
251
Figure 4 Consensus maximum parsimony trees derived from alignments of Gliarin, Macrolin, and
Filarin with other sequences from the IF superfamily. (A) Consensus tree based on an alignment
with all gaps removed (leaving 274 residues) of the rod domains of IF proteins from nuclear,
neuronal, and nonneuronal cytoplasmic invertebrate IFs, as well as representative sequences from
the six classes (I–VI) of vertebrate IFs. The tree is rooted using sequences from three pseudocoelomate invertebrate IFs (nematode ifA3, ifC1, and cif) as an outgroup. (B) Consensus tree based on
an alignment with all gaps removed (leaving 542 residues) of the entire sequence of neuronal and
nonneuronal cytoplasmic invertebrate IFs. The tree is rooted using the nematode sequence ifA3 as
an outgroup. (A,B) the bootstrap 50% majority rule consensus of 1000 maximum parsimony trees
is depicted with associated bootstrap support values. The following IF database sequences were used
[corresponding top to bottom to the list in (A)]: P22488, S01294, S24545, AF101065, S69003,
AF101064, Q01240, X86347, AF101063, S43428, Q03427, P08928, P02546, P02545, P13648,
P09010, 998562, S42257, P20152, P0860, P24790, P48675, U59167, P23239, P21807, P03995,
P08553, P12036, P35617, P19013, P50446, P05787, P02535, Q99456, P21263, P48681, S46327,
AF047657, X070836.
teristic structure with nuclear lamins consisting of six
extra heptads in coil 1B as compared to cytoplasmic
vertebrate IFs which lack these sequences (Weber et
al., 1989; Fuchs and Weber, 1994). This has led to the
proposal that all IFs have evolved from a common
nuclear lamin-like predecessor (Osborn and Weber,
1986; Dodemont et al., 1990). However, our phylogenetic analysis based on maximum parsimony suggests an alternative hypothesis where the original IF
progenitor was a cytoplasmic IF which gave rise to
two monophyletic clades: one comprising all coelomate protostomic cytoplasmic IFs and one comprising
all nuclear lamins as well as the vertebrate cytoplasmic IFs. In this scenario, vertebrate and perhaps all
deutorostomic cytoplasmic IFs may have evolved
from a progenitor shared with nuclear lamins by los-
ing the six heptad repeats in coil 1B (Riemer et al.,
1998) in an event taking place sometime after the
monophyletic clade comprising the coelomate protostomic cytoplasmic IFs had formed. However, it
should be pointed out that at the level of the present
analysis, the ancestral relationship between the major
clades could not definitively be resolved. Regardless,
an important implication of the topology of the phylogenetic trees is that the presence of neuron- and glial
cell–specific IFs in coelomate protostomes as well as
in vertebrates is not of monophyletic origin, but rather
represents convergent evolution and appears to have
arisen independently. This is in contrast to the nuclear
lamins from both protostomes and deutorostomes
which form a distinct monophyletic clade. This analysis was based on an alignment of the rod domains
252
Xu et al.
from which all gaps including that derived from the
lack of the six coil 1B heptad repeats in vertebrate
cytoplasmic IFs were removed, eliminating any possible bias on the results of the presence or absence of
these sequences in a given IF. Moreover, in all our
consensus trees representatives for each of the six
classes of vertebrate IFs were grouped together in
monophyletic clades according to their type, supporting the validity of the topology of the trees. Thus,
maximum parsimony analysis based on the amino
acid sequence promises to provide a valuable framework for resolving issues of IF gene evolution and
how this may relate to the origin of the different
intron/exon patterns of the various IF classes (Dodemont et al., 1990).
It has been suggested that the complexity of invertebrate IFs within the nervous system may be less than
that seen for vertebrates, and that, for example, the
splice variants of the squid IFs are an alternative way
of generating IF functional diversity (Szaro et al.,
1991; Way et al., 1992). This mechanism could function as an adaptation to the relatively smaller genome
size of invertebrates as compared to vertebrates,
where the three major neurofilament components are
encoded by distinct genes (Szaro et al., 1991). However, in this report we show that at least two glial cell
specific IFs, Gliarin and Macrolin, are present within
an invertebrate nervous system. In addition, the phylogenetic analysis clearly demonstrates that leech Filarin is not a homolog of the squid brain IF, indicating
that several neuron-specific IFs may exist in invertebrates as well. Furthermore, the mAB IFA, which
cross-reacts with IF proteins in several invertebrate
and vertebrate species (Pruss et al., 1981), recognizes
at least five prominent protein bands on immunoblots
of leech central nervous system proteins (McKay et
al., 1984), suggesting that still more IFs await identification. It will be interesting to further investigate the
diversity of IFs within the invertebrate nervous system and to determine what role this diversity may play
in maintaining the shape and functional integrity of
individual neuron and glial cells.
The authors thank Dr. R. McKay for help with the
monoclonal antibody fusion, Dr. J. Nicholls and Dr. T.
Lüthi for their generous gift of the G39 antibody, Dr. G.
Naylor for help and advice on the phylogenetic analysis, and
Dr. R. Stewart and Dr. D. Neely for advice on the G39
antibody. They also thank Anna Yeung for expert technical
assistance, as well as Dr. Paul Kapke at the Iowa State
University Hybridoma Facility for help with maintaining
the monoclonal antibody lines. This work was supported by
NIH Grant NS 28857 (JJo), NSF Grant 9724064 (JJe), and
an undergraduate research assistantship (BB) from the
Howard Hughes Medical Institute Education Initiative to
Iowa State University. This is Journal Paper No. J-18159 of
the Iowa Agriculture and Home Economics Experiment
Station, Ames, Iowa, Project No. 3371, and was supported
by Hatch Act and State of Iowa funds.
REFERENCES
Adjaye J, Plessmann U, Weber K, Dodemont H. 1995.
Characterization of neurofilament protein NF70 mRNA
from the gastropod Helix aspersa reveals that neuronal
and non-neuronal intermediate filament proteins of cerebral ganglia arise from separate lamin-related genes.
J Cell Sci 108:3581–3590.
Albers K, Fuchs E. 1992. The molecular biology of intermediate filament proteins. Int Rev Cytol 134:243–279.
Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ.
1990. Basic local alignment search tool. J Mol Biol
215:403– 410.
Bovenschulte M, Riemer D, Weber K. 1995. The sequence
of a cytoplasmic intermediate filament (IF) protein from
the annelid Lumbricus terrestis emphasizes a distinctive
feature of protostomic IF proteins. FEBS Lett 360:223–
226.
Coggeshall RE, Fawcett DW. 1964. The fine structure of the
central nervous system of the leech, Hirudo medicinalis.
J Neurophysiol 27:229 –289.
Devereux J, Haekerli P, Smithies O. 1984. A comprehensive set of sequence programs for the VAX. Nucl Acids
Res 12:387–395.
Dodemont H, Riemer D, Ledger N, Weber K. 1994. Eight
genes and alternative RNA processing pathways generate
an unexpectedly large diversity of cytoplasmic intermediate filament proteins in the nematode Caenorhabditis
elegans. EMBO J 13:2625–2638.
Felsenstein J. 1985. Confidence limits on phylogenies: an
approach using the bootstrap. Evolution 39:783–791.
Flaster MS, Zipser B. 1987. The macroglial cells of the
leech are molecularly heterogeneous. J Neurosci Res
17:176 –183.
Fuchs E, Cleveland DW. 1998. A structural scaffolding of
intermediate filaments in health and disease. Science 279:
514 –519.
Fuchs E, Weber K. 1994. Intermediate filaments: structure,
dynamics, function, and disease. Annu Rev Biochem
63:345–382.
Harlow E, Lane E. 1988. Antibodies: a laboratory manual.
Cold Spring Harbor, NY: Cold Spring Harbor Laboratory
Press.
Huang Y, Jellies J, Johansen KM, Johansen J. 1997. Differential glycosylation of Tractin and LeechCAM, two
novel Ig-superfamily members, regulates neurite extension and fascicle formation. J Cell Biol 138:143–157.
Johansen KM, Johansen J. 1995. Filarin, a novel invertebrate intermediate filament protein present in axons and
perikarya of developing and mature leech neurons. J Neurobiol 27:227–239.
Glial Cell–Specific Intermediate Filaments
Laemmli UK. 1970. Cleavage of structural proteins during
assembly of the head of bacteriophage T4. Nature 227:
680 – 685.
Lendahl U, Zimmerman LB, McKay RDG. 1990. CNS stem
cells express a new class of intermediate filament protein.
Cell 44:283–292.
Lüthi TE, Brodbeck DL, Jenö P. 1994. Identification of a 70
kD protein with sequence homology to squid neurofilament protein in glial cells of the leech CNS. J Neurobiol
25:70 – 82.
Macagno ER. 1980. Number and distribution of neurons in
leech segmental ganglia. J Comp Neurol 190:283–302.
McKay R, Johansen J, Hockfield S. 1984. Monoclonal
antibody identifies a 63,000 dalton antigen found in all
central neuronal cell bodies but in only a subset of axons
in the leech. J Comp Neurol 226:448 – 455.
Morrissey AM, McGlade-McCulloh E. 1988. Development
of identified glia that ensheathe axons in Hirudo medicinalis. J Neurosci Res 21:513–520.
Muller KJ, Nicholls JG, Stent GS. 1981. Neurobiology of
the leech. Cold Spring Harbor, NY: Cold Spring Harbor
Laboratory Press. pp. 320.
Osborn M, Weber K. 1986. Intermediate filament proteins:
a multigene family distinguishing major cell lineages.
Trends Biochem Sci 11:469 – 472.
Pruss RM, Mirsky R, Raff MC, Thorpe R, Dowdling AJ,
Anderton BH. 1981. All classes of intermediate filaments
share a common antigenic determinant defined by a
monoclonal antibody. Cell 27:419 – 428.
Riemer D, Karabinos A, Weber K. 1998. Analysis of eight
cDNAs and six genes for intermediate filament (IF) proteins in the cephalochordate Branchiostoma reveals differences in the IF multigene families of lower chordates
and vertebrates. Gene 211:361–373.
Sambrook J, Fritsch EF, Maniatis T. 1989. Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor,
NY: Cold Spring Harbor Laboratory Press.
253
Steinert PM, Parry DAD. 1985. Intermediate filaments:
conformity and diversity of expression and structure.
Annu Rev Cell Biol 1:41– 65.
Steinert PM, Roop DR. 1988. Molecular and cellular biology of intermediate filaments. Annu Rev Biochem 57:
593– 625.
Swofford DL. 1993. Phylogenetic analysis using parsimony. Champaign, IL: Illinois Natural History Survey.
Szaro BG, Pant HC, Way J, Battey J. 1991. Squid low
molecular weight neurofilament proteins are a novel class
of neurofilament protein. J Biol Chem 266:15035–15041.
Tomarev SI, Zinovieva RD, Piatigorsky J. 1993. Primary
structure and lens-specific expression of genes for an
intermediate filament protein and a b-tubulin in cephalopods. Biochim Biophys Acta 1216:245–254.
Towbin H, Staehelin T, Gordon J. 1979. Electrophoretic
transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc Natl
Acad Sci USA 76:4350 – 4354.
Way J, Hellmich MR, Jaffe H, Szaro B, Pant HC, Gainer H,
Battey J. 1992. A high-molecular-weight squid neurofilament protein contains a lamin-like rod domain and a tail
domain with Lys-Ser-Pro repeats. Proc Natl Acad Sci
USA 89:6963– 6967.
Weber K, Plessman U, Dodemont H, Kossmagk-Stephan K.
1988. Amino acid sequences and homopolymer-forming
ability of the intermediate filament proteins from an invertebrate epithelium. EMBO J 7:2995–3001.
Weber K, Plessman U, Ulrich W. 1989. Cytoplasmic intermediate filament proteins of invertebrates are closer to
nuclear lamins than are vertebrate intermediate filament
proteins; sequence characterization of two muscle proteins of a nematode. EMBO J 8:3221–3227.
Zipser B, McKay R. 1981. Monoclonal antibodies distinguish identifiable neurones in the leech. Nature 289:549 –
554.